Close

Glossary

Term Definition
A/B Test An A/B test compares metrics between 2 groups of users. One group is the control, the other is shown a different version of a product. After enough time has passed for meaningful results to be garnered (generally at least 1 week), various metrics are compared to see which group performed better. A/B tests can be run for more than 2 groups, but care must be taken to ensure that the groups are both random and equal to avoid outliers affecting the final result. In the case that there are more than 2 groups, all groups will be compared to the control.
Accumulated Omniata natively tracks several aspects of users for purposes of creating targeted Campaigns via the use of Segments. Three of these default fields, which are all System User Fields, are Accumulated Revenue, Accumulated Session Count, and Accumulated Session Time. While the primary purpse of these fields is for the building of Segments, they can also be used within a Reporting Table, with one important caveat: all accumulated metrics must be used as Dimensions, not Measures. If these fields are used as Measures, the data will not be aggregated correctly.
Acquirer Acquirer is a Data App that allows visualizations of data related to advertising/user acquisition campaigns across various ad networks.
Acquisition Acquisition is a category of metrics related to how new users come into an ecosystem, and what the users have done since then.
Acquisition Date The day that a user comes into an ecosystem, also called "Day Zero" or "D0".
Activity Date The day that an action occurs.
Android Google's operating system for mobile devices such as phones and tablets
Annotation Omniata allows users to add comments to charts and graphs to draw attention to various data points. For instance, an annotation when a new feature was launched would help explain any sudden spikes or valleys. These annotations persist until they are deleted. To add an annotation, hover over the data point you would like to annotate, and click once.
API An Application Programming Interface (API) specifies the inputs and outputs necessary for a software component to work. Omniata's APIs allow close integration with various operating systems.
API Key A unique identifier that allows events to be parsed. Omniata uses unique API keys for each custom data model to scope which data will be included. Please note that Standard Metrics data models are not scoped by API key, and will pull all data that is sent.
Campaign Campaigns can be promotional campaigns, game configuration based A/B test, push notification campaigns, email campaigns, etc. For the Omniata platform, a Campaign is the culmination of specifying a piece of Content to be delivered to a group of users in a specific way, with a set schedule.
Channel A channel is used to provide targeted content to users.
Churn The likelihood that a user will leave an application and never return. Churn can be calculated in a variety of ways, from simplistic, retention-based metrics to advanced statistical modeling.
Confidence Interval A percentage value indicating the degree of trust we have that a hypothesis is correct. In Omniata, we use this for the results of A/B tests, which are generated by a Z test. A value of 95% or greater can be trusted as is, a value between 90-95% can be trusted directionally, but not the actual value, and a value beneath 90% should not be trusted.
Content Content delivered to users either via the application (promotions, A/B test) or via third party (push notifications, email). The Content is delivered in JSON format.
Content Type Content Type defines the structure of Content, such as what fields are available and how text is displayed.
Conversion Conversion refers to a user's movement between steps in a funnel. These movements may be from clicks to installs in UA, from non-payer to payer in Monetization, or from one part of a game to another. It is calculated by dividing the number of users who successfully completed the action by the total cohort size.
Counter A counter is a type of event field that is incremented each time a specific action occurs. Examples would be sessions or payments. It can also count distinct values during a day, such as distinct user IDs.
CRM Customer Relationship Management
Custom Metrics Custom Metrics data models are uniquely different from Standard Metrics. Custom Metrics data models are fully customizable in regards to types of events sent and analyzed, widgets, and reporting tables, and also offer access to Omniata's Engager.
Custom User Attribute A part of a user's state that is separate from that covered by the user_state. Examples could include a user's level within a game, or the last item they purchased.
Custom User Field One of the 4 types of events in a data model. Custom User fields are stored on a per-user basis, and are not inherent parts of the user state like System User fields. These are only available in Custom Metrics data models. Modifying an existing custom user field, or adding one to a table, will only apply the change on a going-forward basis. To apply the changes to historical data, the table must be reprocessed.
Dashboard A collection of widgets with associated filters with a unique URL. It is a good idea to limit the Data Source for a dashboard's widgets to one specific Reporting Table to avoid filtering issues.
Dashboard Group A collection of dashboards, often with a specific theme (user acquisition, monetization, engagement, progression, etc.)
Data Model A data model is a collection of reporting tables with associated dashboards and widgets. There are 2 basic types of data models. Standard Metrics data models pull data from all events sent, and have a predefined set of widgets, dashboards, dashboard groups, and event types, and jobs cannot be run against the underlying tables. Custom Metrics data models are fully customizable by the user, and unique to specific API keys.
Event Console The Event Console allows developers to monitor the event flow of a test device during integration or when updating an application. The console ensures that the event flow matches the data schema associated with a Data App.
Device Token Used for sending push notifications, each specific device has a token associated with it. A user's device token is mandatory to send push notifications.
Device Type A field that indicates the type of device used to interact with an app or program. Examples include iPhone 5S, HTC Evo, Samsung Galaxy S4, or more generic names like mobile and tablet.
Dimension A field in a reporting table which is not meant to be aggregated. Examples include dates, names, or countries.
Engagement One of the 3 main types of analytics, along with Acquisition and Monetization. Engagement metrics relate to how a user interacts with a particular app or program.
Engager The Engager is one of the 3 main parts of the Omniata platform, along with the Analyzer and the Acquirer. The Engager allows clients to interact with their users via push notifications, email, targeted content, and A/B testing.
Environment Defines the schema that the underlying queries pull data from. The data for each environment is parsed via the API key the events were sent with. Default environments are Production, where all actual data should go, and Staging and Development, which can be used for testing purposes. The environment is indicated by the orange button in the top right of a dashboard. Any jobs run will generate data in all environments available unless scoped to a particular environment.
Event The building block of all data in Omniata. Events are unique pieces of data containing information from users of an app or program, and must contain the event type, API key, and a unique user ID.
Event Based Event based data models are another name for Custom Metrics data models. They are fully customizable as far as KVPs go, as well as all reporting tables, dashboards, and widgets.
Event Field An event field is data from a KVP within an event, which is not saved or associated to a specific user. Examples would include the total number of users, or number of events. Modifying an existing event field, or adding one to a table, will only apply the change on a going-forward basis. To apply the changes to historical data, the table must be reprocessed.
Event Type There are 4 types of data fields in Omniata. Event Fields are the actual data points from the KVPs in an event. System User fields are pulled from the user state. Custom User Fields are KVPs from the events that are saved on a per-user basis. Table Fields are virtual fields that appear as SQL statements.
Experience A particular version of an app or program, or different interaction with the app, which is shown to users in an experiment. Generally, one experience should be the control group, who are shown the existing version, and another will be the test group, who see a modified version.
Experiment The name for a particular A/B test, such as "Faster Progression" or "Sale Prices". An experiment is made up of 2 or more experiences, which are shown to different users. The metrics from each group are then compared to see which group performed best.
Filter Filters change the WHERE and HAVING clauses in the queries that power the widgets in a dashboard. Examples would include specifying a date range or a country or group of countries. Filters can be set to be visible to users or hidden (applied without the Omniata user's knowledge), and default values can also be set. It is a best practice to ensure that each dashboard only pulls from a single reporting table to ensure that the filters are applied correctly to all widgets shown.
Funnel A funnel shows progression through a set of ordered steps. On the acquisition side, this may include view to click to install, and on the in-game side it may include progression through various levels of a game, or use of various parts of a program.
General Metrics (GMET) GMET refers to the reporting tables that power the Standard Metrics data models. These tables cannot be modified, nor can jobs be run against them. Each day's data is aggregated and then appended to the appropriate table.
IDFA ID For Advertisers, Apple's unique identifier for a specific iOS device to allow attribution. The caveat to using this is that users may reset their IDFA at any time.
iOS The operating system powering Apple's mobile devices, often referred to along with the version, such as iOS 7, iOS 8.0.2, etc.
Job A custom processing of a reporting table over a certain date range, which will generate a set of data. These are run either nightly or on-demand, and only work in Custom Metrics data models.
Key Value Pair (KVP) KVPs are part of each individual event. The key is what the system looks up, which will also be the name of an associated event field. The value is what comes after that. For example, in "api_key=1234abcd&uid=zxy5678", the keys are "api_key" and "uid" and the associated values would be "1234abcd" and "zxy5678", respectively.
Lifetime Value (LTV) Lifetime Value is the total projected spend for a user between the time they enter an ecosystem and when they have ceased all activity. Omniata predicts LTV via measuring either Monetization or Retention and projecting either 180 or 360 days of data.
Lookup Table Lookup Tables map values from raw data to the data in the reporting tables. Examples of lookup tables would be to create mappings between country codes and country names, between distinct ages and age ranges, or any other value that needs to be transformed.
Machine Name The internal name of a field or reporting table. This is not the same as the SQL name.
Measure A field in a reporting table that is mean to be aggregated. Examples include revenue and counts of users.
Message Message is a combination of Content, Segment, Campaign, and Experience. The message is requested from the Channel API every time a user logs in and there is an ongoing Campaign.
Monetization One of the 3 main types of analytics, along with Acquisition and Engagement. Monetization metrics answer the what, where, and when as related to users spending money.
MySQL/Postgres Data Model This data model is specifically designed to connect to an external database, in which the data does not exist on the Omniata servers.
Nightly Processing Every night, all the data generated during the day is aggregated and appended to all tables within Omniata. It will generally take a few hours from start to finish.
On-Demand Processing Another term for running a job. If a particular job will take too long, Omniata can also process data as part of the nightly to limit downtime.
Package A group of reporting tables, dashboard groups, dashboards, and widgets, all with a particular theme that may be installed as part of one's panel.
Panel The UI that an organization uses to interact with Omniata.
Platform Platform is a more generic version of device, which is more granular. Examples include iOS, Android, or Windows.
Postback A postback is data sent after a user clicks on an ad, which is used for attribution.
Project A project is a specific source of data, which can be as granular as desired. For instance, projects can be games, or game/platform combinations.
Push Notification Omniata's Engager allows the sending of push notifications, customized messages that are delivered to a user's mobile device outside of the specific app that sends it.
Raw Data The raw data is the unaggregated events that make up all the metrics in the Omniata platform. This data can be accessed via a daily S3 upload if desired.
Reporting Table A group of dimensions and measures that create an underlying SQL table, which widgets query to display data.
REST API A term used to receive and send REST formatted data through APIs.
Retained Days A numeric value showing the difference between a user's Activity Date (when an action ocurred) and a user's Acquisition Date (when a user first sent an event to Omniata). Used for retention metrics.
Retention Retention is a measurement of how many users who installed on a specific day returned later. The number of users who return on a given day after install is divided by the total number of users in that install cohort. Omniata measures retention on a "day of" basis, meaning that "Day 7 Retention" refers to the number of people who came back exactly 7 days after installing. This is as opposed to measuring the number of users who came back within 7 days of installing, which is a separate measure.
Salt When creating A/B tests, a salt is a random value that determines what group a user is assigned to. Randomness is important to avoid bias in the groups.
Scope Scope refers to what particular events are included, either in a data model or reporting table. Custom Metrics data models are scoped by API keys, so only events with the defined keys will be included. Standard Metrics data models are not scoped by API keys. For reporting tables, scope refers to what events are used to create the dimensions and measures. The default is to have all events flow in, but by scoping to a particular event, reporting tables can be tailored to specific analyses. For instance, a reporting table scoped to om_revenue events will only include monetization data, which may be useful for forecasting revenue.
SDK A Software Development Kit standardizes the methodology by which events are sent from a client to Omniata. SDK integrations are the most robust and encouraged for all clients.
Segment A group of users who are targeted by specific attributes, such as payers, users from Europe, users who have reached a certain level, etc. Segments can be used for experiments as well as the targets of content and push notifications.
Session A session is each instance of a user interacting with an app or program. Sessions start upon receipt of the om_load event, and end when there has been a 30-minute window without any events sent.
SQL Name The actual name of a table in SQL, which would be used to write a query.
Standard Metrics A package available from Omniata geared towards acquisition, engagement, and monetization metrics. This package is versatile and the metrics within can be applied to any online program or app.
Statistical Significance A percentage value indicating the chance that the difference between two metrics can be explained as randomness.
System User Attribute Information contained in the user state, referenced by System User fields.
System User Field One of the 4 types of data fields in Omniata. System User fields are created by default in the Omniata panel and display information from the user state. For information not contained in the user state, Custom User fields should be used, except in very narrow use cases. Modifying an existing System User field, or adding one to a table, will only apply the change on a going-forward basis. To apply the changes to historical data, the table must be reprocessed.
Table Field A virtual field that doesn't exist in the actual data, but only as a statement within the SELECT clause of a query. These use SQL syntax and are only applied when a query is run, so adding a new table field to a reporting table will not require reprocessing.
Tier A tier is a range of values that are mapped together, such as ages 18-35, or countries that speak English. These are generally used in lookup tables when the raw data is at a level of granularity that is unneeded, or if there is a need to group users together for analysis.
Transience Due to the way Omniata's data is processed, when a user moves between 2 states (such as payer and non-payer), this can create 2 rows of data in the underlying SQL table, one with the old value, and one with the new value. Transient dimensions can throw off calculations or mis-allocate measurements if they are not accounted for.
UID A unique identifier for a specific user. UIDs are used to calculate DAU and associated metrics, and are a necessary component of every event sent to Omniata.
Unity A cross-platform engine often used for games. Omniata supports integration with Unity-based games.
Unix Timestamp The number of seconds since January 1st, 1970 (UTC time zone). There is a variety of ways to convert this to a human-readable value via Excel or SQL formulas.
User State A per-user record of all actions tracked by Omniata. This shows a user's complete history as it relates to engagement, monetization, and acquisition.
Widget A widget shows the results of a query against a user's data, displayed either graphically in a chart, or the actual values in a table. The underlying queries can be changed via filters on the page, which in essence change the WHERE and HAVING clauses of the query.
X-Forwarded-For The X-Forwarded-For (XFF) HTTP header field is a de facto standard for identifying the originating IP address of a client connecting to a web server through an HTTP proxy or load balancer. This is an HTTP request header that was introduced by the Squid caching proxy server's developers.
Z Test A Z test is a statistical operation performed on 2 proportions or averages. It allows you to determine if the differences between the two values are "statistically significant", or in other words, if the values are different for a reason besides random noise.

This article was last updated on November 11, 2015 22:12. If you didn't find your answer here, search for another article or contact our support to get in touch.

Contact Sales

Add Phone Number